Thanks James.
I'm actually trying to debug this on a new server with the latest stable release of everything installed - doesn't matter if I screw things up as is highly likely since I'm climbing a steep learning curve on how runJobs.php works :-)
I'm making progress but have hit another problem which is obviously a function of removed php or MySQL commands/functions which were only deprecated in the earlier versions on our production server.
I believe I am right in saying that to properly delete all records in the job table using mySQL commands does NOT require deletion of records in the corresponding index tables - at least having done so, followed by some page edits to see their effect on the queue, there does not appear to have been any damage to wiki functionality at all. If anyone knows differently I'd appreciate a pointer.
I hope it does turn out to be a SMW 2.0 problem but from experience so far I think it more likely to be a circular reference in either a page redirect or something similar in one of our horrendously complicated SMW templates.
Having emptied the job table, rebuilt the SMW store and run both rebuildData and Update, I am now doing non-functional edits on the suspect templates and closely monitoring their effect on the queue - very educational it is too.
I'll keep the list posted and meantime, I'd welcome suggestions from anyone else too.
Peter P
On 27/04/2015 12:30, James Montalvo wrote:
So 24 hours ago you were at 80000 jobs and now you're at 69000? This sounds like a problem that was fixed in SMW 2.1. See [0]. I recommend upgrading. Once runJobs has run it's course I'd run the rebuildData script anyway just to confirm you're in good shape.
Since we upgraded to SMW 2.1 we haven't had runJobs issues anymore.
--James